Preference alignment has become a crucial component in enhancing the performance of Large Language Models (LLMs), yet its impact in Multimodal Large Language Models (MLLMs) remains comparatively underexplored. Similar to language models, MLLMs for image understanding tasks encounter challenges like hallucination. In MLLMs, hallucination can occur not only by stating incorrect facts but also by producing responses that are inconsistent with the image content. A primary objective of alignment for MLLMs is to encourage these models to align responses more closely with image information. Recently, multiple works have introduced preference datasets for MLLMs and examined different alignment methods, including Direct Preference Optimization (DPO) and Proximal Policy Optimization (PPO). However, due to variations in datasets, base model types, and alignment methods, it remains unclear which specific elements contribute most significantly to the reported improvements in these works. In this paper, we independently analyze each aspect of preference alignment in MLLMs. We start by categorizing the alignment algorithms into two groups, offline (such as DPO), and online (such as online-DPO), and show that combining offline and online methods can improve the performance of the model in certain scenarios. We review a variety of published multimodal preference datasets and discuss how the details of their construction impact model performance. Based on these insights, we introduce a novel way of creating multimodal preference data called Bias-Driven Hallucination Sampling (BDHS) that needs neither additional annotation nor external models, and show that it can achieve competitive performance to previously published alignment work for multimodal models across a range of benchmarks.
Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；Seoul National University（首尔国立大学）；University of Washington（华盛顿大学）
Large language models readily adapt to novel settings, even without task-specific training data. Can their zero-shot capacity be extended to multimodal inputs? In this work, we propose ESPER which extends language-only zero-shot models to unseen multimodal tasks, like image and audio captioning. Our key novelty is to use reinforcement learning to align multimodal inputs to language model generations without direct supervision: for example, in the image case our reward optimization relies only on cosine similarity derived from CLIP, and thus requires no additional explicitly paired (image, capt…
Single-to-mix Modality Alignment with Multimodal Large Language Model for Document Image Machine Translation
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Zhongke Fanyu Technology Co., Ltd（中科泛语科技有限公司）
PivotMerge: Bridging Heterogeneous Multimodal Pre-training via Post-Alignment Model Merging
State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences（多模态人工智能系统国家重点实验室，中国科学院自动化研究所）；Pengcheng Laboratory（鹏城实验室）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；Data Science and Artificial Intelligence Research Institute, China United Network Communications Group Co., Ltd.（中国联合网络通信集团有限公司数据科学与人工智能研究院）
MAGE: Multimodal Alignment and Generation Enhancement via Bridging Visual and Semantic Spaces
Global Tone Communication Technology Co., Ltd.（全球 tone 通信技术有限公司）；Faculty of computing, Harbin Institute of Technology（哈尔滨工业大学计算机学院）；School of Computer Science and Technology, Beijing Jiaotong University（北京交通大学计算机科学与技术学院）
LLMs Meet Multimodal Generation and Editing: A Survey
The Hong Kong University of Science and Technology（香港科技大学）；The Chinese University of Hong Kong（香港中文大学）；Tsinghua University（清华大学）；Tencent AI Lab（腾讯AI实验室）
ShizhenGPT: Towards Multimodal LLMs for Traditional Chinese Medicine
The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））；Freedom AI；National Health Data Institute, Shenzhen（深圳国家健康数据研究院）；Shenzhen Research Institute of Big Data（深圳大数据研究院）；Guangdong University of Technology（广东工业大学）
mChartQA: A universal benchmark for multimodal Chart Question Answer based on Vision-Language Alignment and Reasoning
Shenyang Institute of Computing Technology, Chinese Academy of Sciences（中国科学院沈阳计算技术研究所）；Beijing Wenge Technology Co., Ltd（北京文格科技有限公司）；University of Chinese Academy of Sciences（中国科学院大学）
SAE-V: Interpreting Multimodal Models for Enhanced Alignment
Institute for AI, Peking University, Beijing, China（人工智能研究院，北京大学，北京，中国）；State Key Laboratory of General Artificial Intelligence, Institute for AI, Peking University, Beijing, China（通用人工智能国家重点实验室，人工智能研究院，北京大学，北京，中国）
EAMA : Entity-Aware Multimodal Alignment Based Approach for News Image Captioning
Wangxuan Institute of Computer Technology, Peking University（北京大学王选计算机研究所）；School of Electronics Engineering and Computer Science, Peking University（北京大学电子工程与计算机科学学院）
Large Language Model Guided Progressive Feature Alignment for Multimodal UAV Object Detection
School of Artificial Intelligence, Anhui University（安徽大学人工智能学院）；School of Computer Science and Technology, Anhui University（安徽大学计算机科学与技术学院）；University of Science and Technology of China（中国科学技术大学）
Shanghai Jiao Tong University（上海交通大学）；UNC Chapel Hill（北卡罗来纳大学教堂山分校）；University of California, Berkeley（加州大学伯克利分校）
Phrase localization is a task that studies the mapping from textual phrases to regions of an image. Given difficulties in annotating phrase-to-object datasets at scale, we develop a Multimodal Alignment Framework (MAF) to leverage more widely-available caption-image datasets, which can then be used as a form of weak supervision. We first present algorithms to model phrase-object relevance by leveraging fine-grained visual representations and visually-aware language representations. By adopting a contrastive objective, our method uses information in caption-image pairs to boost the performance …
